Updated on 2024/01/02

写真a

 
OKITA Tsuyoshi
 
Scopus Paper Info  
Total Paper Count: 0  Total Citation Count: 0  h-index: 10

Citation count denotes the number of citations in papers published for a particular year.

Affiliation
Faculty of Computer Science and Systems Engineering Department of Artificial Intelligence
Job
Associate Professor
E-mail
メールアドレス
Laboratory
680-4 Kawazu, Iizuka-shi, Fukuoka
External link

Research Interests

  • artificial intelligence

  • machine learning

Research Areas

  • Informatics / Soft computing

Post Graduate Education

  • 2012.03   ダブリン市大学(Dublin City University)   コンピュータ学科   Doctoral Program   Completed   Ireland

Degree

  • Dublin City University  -  Doctor of Philosophy   2012.03

Biography in Kyutech

  • 2022.04
     

    Kyushu Institute of Technology   Faculty of Computer Science and Systems Engineering   Department of Artificial Intelligence   Associate Professor  

  • 2019.04
    -
    2022.03
     

    Kyushu Institute of Technology   Faculty of Computer Science and Systems Engineering   Department of Artificial Intelligence   Specially Appointed Associate Professor  

  • 2018.04
    -
    2019.03
     

    Kyushu Institute of Technology   Faculty of Engineering   Department of Basic Sciences   Specially Appointed Lecturer  

  • 2017.01
    -
    2018.03
     

    Kyushu Institute of Technology   Faculty of Engineering   Department of Basic Sciences  

Papers

  • Towards LLMs for Sensor Data: Multi-Task Self-Supervised Learning Reviewed

    Tsuyoshi Okita, Kosuke Ukita, Koki Matsuishi, Masaharu Kagiyama, Kodai Hirata, Asahi Miyazaki

    Adjunct Proceedings of the 2023 ACM International Joint Conference on Pervasive and Ubiquitous Computing & the 2023 ACM International Symposium on Wearable Computing (UbiComp/ISWC 2023)   2023.10

     More details

    Authorship:Lead author, Corresponding author   Language:English   Publishing type:Research paper (international conference proceedings)

    Mexico  

  • 11th International Workshop on Human Activity Sensing Corpus and Applications (HASCA) Reviewed International journal

    Murao K., Enokibori Y., Gjoreski H., Lago P., Okita T., Siirtola P., Hiroi K., Scholl P.M., Ciliberto M., Urano K.

    UbiComp/ISWC 2023 Adjunct - Adjunct Proceedings of the 2023 ACM International Joint Conference on Pervasive and Ubiquitous Computing and the 2023 ACM International Symposium on Wearable Computing   773 - 776   2023.10

     More details

    Language:English   Publishing type:Research paper (international conference proceedings)

    DOI: 10.1145/3594739.3605106

    Scopus

    Other Link: https://www.scopus.com/inward/record.uri?partnerID=HzOxMe3b&scp=85175476451&origin=inward

  • Towards LLMs for Sensor Data: Multi-Task Self-Supervised Learning Reviewed International journal

    Okita T., Ukita K., Matsuishi K., Kagiyama M., Hirata K., Miyazaki A.

    UbiComp/ISWC 2023 Adjunct - Adjunct Proceedings of the 2023 ACM International Joint Conference on Pervasive and Ubiquitous Computing and the 2023 ACM International Symposium on Wearable Computing   499 - 504   2023.10

     More details

    Authorship:Lead author, Corresponding author   Language:English   Publishing type:Research paper (international conference proceedings)

    DOI: 10.1145/3594739.3610745

    Scopus

    Other Link: https://www.scopus.com/inward/record.uri?partnerID=HzOxMe3b&scp=85175472929&origin=inward

  • Summary of SHL Challenge 2023: Recognizing Locomotion and Transportation Mode from GPS and Motion Sensors Reviewed International journal

    Wang L., Gjoreski H., Ciliberto M., Lago P., Murao K., Okita T., Roggen D.

    UbiComp/ISWC 2023 Adjunct - Adjunct Proceedings of the 2023 ACM International Joint Conference on Pervasive and Ubiquitous Computing and the 2023 ACM International Symposium on Wearable Computing   575 - 585   2023.10

     More details

    Language:English   Publishing type:Research paper (international conference proceedings)

    DOI: 10.1145/3594739.3610758

    Scopus

    Other Link: https://www.scopus.com/inward/record.uri?partnerID=HzOxMe3b&scp=85175460992&origin=inward

  • Characteristics of Li-Ion Battery at Accelerated C-Rate with Deep Learning Method Reviewed International journal

    Hoque M.A., Hassan M.K., Hajjo A., Okita T.

    Arabian Journal for Science and Engineering   2023.03

     More details

    Authorship:Last author   Language:English   Publishing type:Research paper (scientific journal)

    DOI: 10.1007/s13369-023-08034-x

    Scopus

    Other Link: https://www.scopus.com/inward/record.uri?partnerID=HzOxMe3b&scp=85162231470&origin=inward

  • 2Dアニメーションのフレーム補間に関するサーベイ

    河津水紀, 大北 剛

    第47回IBISML研究会   2022.09

     More details

    Authorship:Last author   Language:Japanese   Publishing type:Research paper (conference, symposium, etc.)

  • 人物実写画を特定の漫画の画風に沿うように変換し画像を出力する技術のサーベイ

    中島崇晴,大北 剛

    第47回IBISML研究会   2022.09

     More details

    Authorship:Last author   Language:Japanese   Publishing type:Research paper (conference, symposium, etc.)

  • 10th International Workshop on Human Activity Sensing Corpus and Applications (HASCA) Reviewed

    Murao K., Enokibori Y., Gjoreski H., Lago P., Okita T., Siirtola P., Hiroi K., Scholl P.M., Ciliberto M., Urano K.

    UbiComp/ISWC 2022 Adjunct - Proceedings of the 2022 ACM International Joint Conference on Pervasive and Ubiquitous Computing and Proceedings of the 2022 ACM International Symposium on Wearable Computers   321 - 323   2022.09

     More details

    Language:English   Publishing type:Research paper (international conference proceedings)

    DOI: 10.1145/3544793.3560377

    Scopus

    Other Link: https://www.scopus.com/inward/record.uri?partnerID=HzOxMe3b&scp=85158983039&origin=inward

  • 脳血腫マーカーの画像パッチのマルチラベル学習

    加藤 舜斗, 河津 水紀, 中島 崇晴, 有村 公一, 飯原 弘二, 大北 剛

    DICOMOシンポジウム   2022.07

     More details

    Authorship:Last author   Language:Japanese   Publishing type:Research paper (conference, symposium, etc.)

  • ハミルトニアンニューラルネットワークの人間行動認識への応用

    豊坂 祐樹, 大北 剛

    DICOMOシンポジウム   2022.07

     More details

    Authorship:Last author   Language:Japanese   Publishing type:Research paper (conference, symposium, etc.)

  • 動きからの構造: 2D画像からの3Dテクスチャの再構築

    大野 友暉, 大北 剛

    DICOMOシンポジウム   2022.07

     More details

    Authorship:Last author   Language:Japanese   Publishing type:Research paper (conference, symposium, etc.)

  • 脳血腫の急成長の予測

    大北剛,中山俊太朗,山本周平,森山幹太,平野北斗,有村公一,飯原弘二

    第12回AIM 合同研究会   2022.03

     More details

    Authorship:Lead author   Language:Japanese   Publishing type:Research paper (conference, symposium, etc.)

  • A Compression-Based Multiple Subword Segmentation for Neural Machine Translation Reviewed International journal

    Keita Nonaka , Kazutaka Yamanouchi , Tomohiro I , Tsuyoshi Okita , Kazutaka Shimada, Hiroshi Sakamoto

    Electronics ( MDPI )   11 ( 7 )   2022.03

     More details

  • 距離画像推定情報を用いた複数人の行動認識

    豊坂祐樹,大北剛

    第72回UBI研究会   2021.11

     More details

    Authorship:Last author   Language:Japanese   Publishing type:Research paper (conference, symposium, etc.)

  • 弱教師ありセマンティックセグメンテーション

    大北剛,平野北斗,森山幹太,有村公一,飯原弘二

    第23回日本知能情報ファジィ学会九州支部学術講演会(ソフト九州)   2021.11

     More details

    Authorship:Lead author   Language:Japanese   Publishing type:Research paper (conference, symposium, etc.)

  • Classification of hematoma marker: classification, object detection, GAN, and semantic segmentation

    OKITA Tsuyoshi, HIRANO Hokuto, ARIMURA Koichi, IIHARA Koji

    JSAI Technical Report, Type 2 SIG ( The Japanese Society for Artificial Intelligence )   2021 ( AIMED-011 )   05   2021.11

     More details

    Authorship:Lead author, Corresponding author   Language:Japanese   Publishing type:Research paper (scientific journal)

    DOI: 10.11517/jsaisigtwo.2021.aimed-011_05

    CiNii Research

  • Locomotion and Transportation Mode Recognition from GPS and Radio Signals: Summary of SHL Challenge 2021 Reviewed

    Wang L., Ciliberto M., Gjoreski H., Lago P., Murao K., Okita T., Roggen D.

    UbiComp/ISWC 2021 - Adjunct Proceedings of the 2021 ACM International Joint Conference on Pervasive and Ubiquitous Computing and Proceedings of the 2021 ACM International Symposium on Wearable Computers   412 - 422   2021.09

     More details

    Language:English   Publishing type:Research paper (international conference proceedings)

    DOI: 10.1145/3460418.3479373

    Scopus

    Other Link: https://www.scopus.com/inward/record.uri?partnerID=HzOxMe3b&scp=85115916446&origin=inward

  • 9th International Workshop on Human Activity Sensing Corpus and Applications (HASCA) Reviewed

    Murao K., Enokibori Y., Gjoreski H., Lago P., Okita T., Siirtola P., Hiroi K., Scholl P.M., Ciliberto M., Urano K.

    UbiComp/ISWC 2021 - Adjunct Proceedings of the 2021 ACM International Joint Conference on Pervasive and Ubiquitous Computing and Proceedings of the 2021 ACM International Symposium on Wearable Computers   281 - 284   2021.09

     More details

    Authorship:Corresponding author   Language:English   Publishing type:Research paper (international conference proceedings)

    DOI: 10.1145/3460418.3479266

    Scopus

    Other Link: https://www.scopus.com/inward/record.uri?partnerID=HzOxMe3b&scp=85115942696&origin=inward

  • Activity Knowledge Graph Recognition by Eye Gaze: Identification of Distant Object in Eye Sight for Watch Activity Reviewed

    Toyosaka Y., Okita T.

    UbiComp/ISWC 2021 - Adjunct Proceedings of the 2021 ACM International Joint Conference on Pervasive and Ubiquitous Computing and Proceedings of the 2021 ACM International Symposium on Wearable Computers   334 - 339   2021.09

     More details

    Authorship:Corresponding author   Language:English   Publishing type:Research paper (international conference proceedings)

    DOI: 10.1145/3460418.3479351

    Scopus

    Other Link: https://www.scopus.com/inward/record.uri?partnerID=HzOxMe3b&scp=85115984916&origin=inward

  • Activity Simulation from Signals Reviewed

    Okita T.

    UbiComp/ISWC 2021 - Adjunct Proceedings of the 2021 ACM International Joint Conference on Pervasive and Ubiquitous Computing and Proceedings of the 2021 ACM International Symposium on Wearable Computers   59 - 60   2021.09

     More details

    Authorship:Lead author   Language:English   Publishing type:Research paper (international conference proceedings)

    DOI: 10.1145/3460418.3479275

    Scopus

    Other Link: https://www.scopus.com/inward/record.uri?partnerID=HzOxMe3b&scp=85115964745&origin=inward

  • Three-Year Review of the 2018–2020 SHL Challenge on Transportation and Locomotion Mode Recognition From Mobile Sensors Reviewed International journal

    Wang L., Gjoreski H., Ciliberto M., Lago P., Murao K., Okita T., Roggen D.

    Frontiers in Computer Science   3   2021.09

     More details

    Language:English   Publishing type:Research paper (scientific journal)

    DOI: 10.3389/fcomp.2021.713719

    Scopus

    Other Link: https://www.scopus.com/inward/record.uri?partnerID=HzOxMe3b&scp=85115915802&origin=inward

  • Summary of the Third Nurse Care Activity Recognition Challenge-Can We Do from the Field Data? Reviewed

    Alia S.S., Adachi K., Hossain T., Le N.T., Kaneko H., Lago P., Okita T., Inoue S.

    UbiComp/ISWC 2021 - Adjunct Proceedings of the 2021 ACM International Joint Conference on Pervasive and Ubiquitous Computing and Proceedings of the 2021 ACM International Symposium on Wearable Computers   428 - 433   2021.09

     More details

    Language:English   Publishing type:Research paper (international conference proceedings)

    DOI: 10.1145/3460418.3479391

    Scopus

    Other Link: https://www.scopus.com/inward/record.uri?partnerID=HzOxMe3b&scp=85115908587&origin=inward

  • 非接触性の人と物体における相互作用検出方法の提案

    豊坂祐樹, 大北剛

    DICOMOシンポジウム   2021.06

     More details

    Authorship:Last author   Language:Japanese   Publishing type:Research paper (conference, symposium, etc.)

  • 脳血腫の分類: セグメンテーションと分類のジョイント学習

    平野北斗, 大北剛

    arxiv   2021.03

     More details

    Authorship:Corresponding author   Language:Japanese   Publishing type:Research paper (scientific journal)

    arXiv

  • Improvement of Human Action Recognition Using 3D Pose Estimation Reviewed

    Adachi K., Lago P., Okita T., Inoue S.

    Smart Innovation, Systems and Technologies   204   21 - 37   2021.01

     More details

    Language:English   Publishing type:Research paper (international conference proceedings)

    While Human Action Recognition (HAR) using motion capture can perform well with high accuracy, it requires a high computational cost for recording and post-processing. To avoid this, we build a HAR system using 3D pose estimation from the single-camera video instead of motion capture. One drawback in this approach is that the performance is considerably dependent on the camera position. This paper investigates how we can use the pose estimate constantly without the effect of camera position even when the camera position in the test data is changed. We augment the data by rotating around the 3D pose estimate to improve the accuracy when using different camera positions in the test data and in the training data. The strategy of augmenting training data shows improvements up to 55.7% in accuracy, compared with the case of 2D pose with no augmentation.

    DOI: 10.1007/978-981-15-8944-7_2

    Scopus

    Other Link: https://www.scopus.com/inward/record.uri?partnerID=HzOxMe3b&scp=85107353674&origin=inward

  • 会議報告:NeurIPS 2020(The 34th Conference on Neural Information Processing Systems) Reviewed

    高瀬 翔, 坂本 陸, 大北 剛

    Journal of the Japanese Society for Artificial Intelligence ( The Japanese Society for Artificial Intelligence )   36 ( 4 )   527 - 531   2021.01

     More details

    Language:English   Publishing type:Article, review, commentary, editorial, etc. (scientific journal)

    DOI: 10.11517/jjsai.36.4_527

    CiNii Article

    Other Link: https://ci.nii.ac.jp/naid/130008060910

  • Summary of the sussex-huawei locomotion-transportation recognition challenge 2020 Reviewed

    Wang L., Gjoreski H., Ciliberto M., Lago P., Murao K., Okita T., Roggen D.

    UbiComp/ISWC 2020 Adjunct - Proceedings of the 2020 ACM International Joint Conference on Pervasive and Ubiquitous Computing and Proceedings of the 2020 ACM International Symposium on Wearable Computers   351 - 358   2020.09

     More details

    Language:English   Publishing type:Research paper (international conference proceedings)

    In this paper we summarize the contributions of participants to the third Sussex-Huawei Locomotion-Transportation (SHL) Recognition Challenge organized at the HASCA Workshop of UbiComp/ISWC 2020. The goal of this machine learning/data science challenge is to recognize eight locomotion and transportation activities (Still, Walk, Run, Bike, Bus, Car, Train, Subway) from the inertial sensor data of a smartphone in a user-independent manner with an unknown target phone position. The training data of a "train"user is available from smartphones placed at four body positions (Hand, Torso, Bag and Hips). The testing data originates from "test"users with a smartphone placed at one, but unknown, body position. We introduce the dataset used in the challenge and the protocol of the competition. We present a meta-analysis of the contributions from 15 submissions, their approaches, the software tools used, computational cost and the achieved results. Overall, one submission achieved F1 scores above 80%, three with F1 scores between 70% and 80%, seven between 50% and 70%, and four below 50%, with a latency of maximum of 5 seconds.

    DOI: 10.1145/3410530.3414341

    Scopus

    Other Link: https://www.scopus.com/inward/record.uri?partnerID=HzOxMe3b&scp=85091885062&origin=inward

  • 8th international workshop on human activity sensing corpus and applications (HASCA) Reviewed

    Murao K., Enokibori Y., Gjoreski H., Lago P., Okita T., Siirtola P., Hiroi K., Scholl P.M., Ciliberto M.

    UbiComp/ISWC 2020 Adjunct - Proceedings of the 2020 ACM International Joint Conference on Pervasive and Ubiquitous Computing and Proceedings of the 2020 ACM International Symposium on Wearable Computers   228 - 231   2020.09

     More details

    Language:English   Publishing type:Research paper (international conference proceedings)

    The recognition of complex and subtle human behaviors from wearable sensors will enable next-generation human-oriented computing in scenarios of high societal value (e.g., dementia care). This will require large-scale human activity corpus and much improved methods to recognize activities and the context in which they occur. This workshop deals with the challenges of designing reproducible experimental setups, running large-scale dataset collection campaigns, designing activity and context recognition methods that are robust and adaptive, and evaluating systems in the real world. We wish to reflect on future methods, such as lifelong learning approaches that allow open-ended activity recognition. This year HASCA will welcome papers from participants to the Third Sussex-Huawei Locomotion and Transportation Recognition Challenge and Second Nursing Activity Recognition Challenge in special sessions.

    DOI: 10.1145/3410530.3414612

    Scopus

    Other Link: https://www.scopus.com/inward/record.uri?partnerID=HzOxMe3b&scp=85091833400&origin=inward

  • Perception of interaction between hand and object Reviewed

    Toyosaka Y., Okita T.

    UbiComp/ISWC 2020 Adjunct - Proceedings of the 2020 ACM International Joint Conference on Pervasive and Ubiquitous Computing and Proceedings of the 2020 ACM International Symposium on Wearable Computers   290 - 295   2020.09

     More details

    Language:English   Publishing type:Research paper (international conference proceedings)

    Action knowledge graphs can play a central role in smart cities, smart homes, robot planning, and so on. This is since both of the subject and the object of the actions can add more meaningful information for the higher-level application than the action alone as a predicate. We built a system that generates the action knowledge graphs from video using deep learning. Especially, we propose an algorithm which perceives the interaction between hand and object by measuring the proximity between them with considering the direction of fingers. We showed that this approach achieves the performance of 83% in accuracy using the Stair Lab data.

    DOI: 10.1145/3410530.3414363

    Scopus

    Other Link: https://www.scopus.com/inward/record.uri?partnerID=HzOxMe3b&scp=85091839586&origin=inward

  • Summary of the 2nd nurse care activity recognition challenge using lab and field data Reviewed

    Alia S.S., Lago P., Adachi K., Hossain T., Goto H., Okita T., Inoue S.

    UbiComp/ISWC 2020 Adjunct - Proceedings of the 2020 ACM International Joint Conference on Pervasive and Ubiquitous Computing and Proceedings of the 2020 ACM International Symposium on Wearable Computers   378 - 383   2020.09

     More details

    Language:English   Publishing type:Research paper (international conference proceedings)

    2nd Nurse Care Activity Recognition Challenge using Lab and Field Data is organized as a part of HASCA workshop and is continuation of Nurse Care Activity Recognition Challenge [7]. We give the description of the dataset and summarize the approaches used by the teams in this Challenge. In this challenge, data collected in both lab and real-world setting is provided to the challenge participants with an aim to bridge the gap between lab and practical field to reduce the workload of the nurses. The challenge was started on May 1, 2020 and continued until July 9, 2020. Accuracy is used as performance metric to evaluate the submissions. The winning team used k-NN classifier and achieved about 22.35% accuracy.

    DOI: 10.1145/3410530.3414611

    Scopus

    Other Link: https://www.scopus.com/inward/record.uri?partnerID=HzOxMe3b&scp=85091886199&origin=inward

  • Translation Between Waves, wave2wave International journal

    Tsuyoshi Okita, Hirotaka Hachiya, Sozo Inoue, Naonori Ueda

    arxiv   2020.07

     More details

    Authorship:Lead author   Language:English   Publishing type:Research paper (scientific journal)

  • 手と物体の相互作用の認識とその知識グラフの生成

    豊坂祐樹, 大北剛

    DICOMOシンポジウム   361 - 367   2020.06

     More details

    Language:Japanese   Publishing type:Research paper (conference, symposium, etc.)

    Virtual   2020.06.30  -  2020.07.02

  • U-netを用いた異常検知による肺炎の検知

    長村 徹, 徳永 旭将, 大北 剛

    DICOMOシンポジウムプロシーディングス   526 - 533   2020.06

     More details

    Language:Japanese   Publishing type:Research paper (conference, symposium, etc.)

  • SHAP値や重要度を用いたモデル解釈性: 包除積分ネットワークとXGBoostの比較

    板橋将之, 本田あおい, 大北剛

    火の国シンポジウムプロシーディングス   2020.03

     More details

    Authorship:Corresponding author   Language:Japanese   Publishing type:Research paper (conference, symposium, etc.)

  • アマゾン商品レビューの感情分析と評価スコアの相関性の検証

    高山なつき, Mario Koppen, 大北剛

    火の国シンポジウムプロシーディングス   2020.03

     More details

    Authorship:Corresponding author   Language:Japanese   Publishing type:Research paper (conference, symposium, etc.)

  • 会議報告:The 33rd Conference on Neural Information Processing Systems(NeurIPS 2019) Reviewed

    大北 剛

    Journal of the Japanese Society for Artificial Intelligence ( The Japanese Society for Artificial Intelligence )   35 ( 3 )   465 - 466   2020.01

     More details

    Language:English   Publishing type:Article, review, commentary, editorial, etc. (scientific journal)

    DOI: 10.11517/jjsai.35.3_465

    CiNii Article

    Other Link: https://ci.nii.ac.jp/naid/130007917827

  • ビデオからの3次元姿勢推定と機械学習を用いた行動認識の試み

    安達康平,大北剛,井上創造

    ソフト九州プロシーディングス   2019.11

     More details

    Language:Japanese   Publishing type:Research paper (conference, symposium, etc.)

  • 不変リスク最小化の考察

    大北剛

    ソフト九州プロシーディングス   2019.11

     More details

    Authorship:Lead author   Language:Japanese   Publishing type:Research paper (conference, symposium, etc.)

  • 混合正規分布のデータセットにおけるAntlion clusteringと既存のクラスタリング手法との比較

    豊坂祐樹,福田亮治,大北剛,宮野英次

    ソフト九州プロシーディングス   2019.11

     More details

    Language:Japanese   Publishing type:Research paper (conference, symposium, etc.)

  • Summary of the Sussex-Huawei locomotion-transportation recognition challenge 2019 Reviewed

    Wang L., Gjoreski H., Ciliberto M., Lago P., Murao K., Okita T., Roggen D.

    UbiComp/ISWC 2019- - Adjunct Proceedings of the 2019 ACM International Joint Conference on Pervasive and Ubiquitous Computing and Proceedings of the 2019 ACM International Symposium on Wearable Computers   849 - 856   2019.09

     More details

    Language:English   Publishing type:Research paper (international conference proceedings)

    In this paper we summarize the contributions of participants to the Sussex-Huawei Transportation-Locomotion (SHL) Recognition Challenge organized at the HASCA Workshop of UbiComp 2019. The goal of this machine learning/data science challenge is to recognize eight locomotion and transportation activities (Still, Walk, Run, Bike, Bus, Car, Train, Subway) from the inertial sensor data of a smartphone in a placement independent manner. The training data is collected with smartphones placed at three body positions (Torso, Bag and Hips), while the testing data is collected with a smartphone placed at another body position (Hand). We introduce the dataset used in the challenge and the protocol for the competition. We present a meta-analysis of the contributions from 14 submissions, their approaches, the software tools used, computational cost and the achieved results. Overall, three submissions achieved F1 scores between 70% and 80%, five with F1 scores between 60% and 70%, five between between 50% and 60%, and one below 50%, with a latency of a maximum of 5 seconds.

    DOI: 10.1145/3341162.3344872

    Scopus

    Other Link: https://www.scopus.com/inward/record.uri?partnerID=HzOxMe3b&scp=85072892002&origin=inward

  • 7th international workshop on human Activity Sensing Corpus and Applications (HASCA) Reviewed

    Murao K., Enokibori Y., Gjoreski H., Lago P., Okita T., Siirtola P., Hiroi K., Scholl P.M., Ciliberto M.

    UbiComp/ISWC 2019- - Adjunct Proceedings of the 2019 ACM International Joint Conference on Pervasive and Ubiquitous Computing and Proceedings of the 2019 ACM International Symposium on Wearable Computers   671 - 673   2019.09

     More details

    Language:English   Publishing type:Research paper (international conference proceedings)

    The recognition of complex and subtle human behaviors from wearable sensors will enable next-generation human-oriented computing in scenarios of high societal value (e.g., dementia care). This will require large-scale human activity corpuses and much improved methods to recognize activities and the context in which they occur. This workshop deals with the challenges of designing reproducible experimental setups, running large-scale dataset collection campaigns, designing activity and context recognition methods that are robust and adaptive, and evaluating systems in the real world. We wish to reflect on future methods, such as lifelong learning approaches that allow open-ended activity recognition. This year HASCA will welcome papers from participants to the Second Sussex-Huawei Locomotion and Transportation Recognition Competition and Open Lab Nursing Activity Recognition Challenge in special sessions.

    DOI: 10.1145/3341162.3347765

    Scopus

    Other Link: https://www.scopus.com/inward/record.uri?partnerID=HzOxMe3b&scp=85072892160&origin=inward

  • Nurse care activity recognition challenge: Summary and results Reviewed

    Lago P., Alia S.S., Takeda S., Mairittha T., Mairittha N., Faiz F., Nishimura Y., Adachi K., Okita T., Charpillet F., Inoue S.

    UbiComp/ISWC 2019- - Adjunct Proceedings of the 2019 ACM International Joint Conference on Pervasive and Ubiquitous Computing and Proceedings of the 2019 ACM International Symposium on Wearable Computers   746 - 751   2019.09

     More details

    Language:English   Publishing type:Research paper (international conference proceedings)

    Although activity recognition has been studied for a long time now, research and applications have focused on physical activity recognition. Even if many application domains require the recognition of more complex activities, research on such activities has attracted less attention. One reason for this gap is the lack of datasets to evaluate and compare different methods. To promote research in such scenarios, we organized the Open Lab Nursing Activity Recognition Challenge focusing on the recognition of complex activities related to the nursing domain. Nursing domain is one of the domains that can benefit enormously from activity recognition but has not been researched due to lack of datasets. The competition used the CARECOM Nurse Care Activity Dataset, featuring 7 activities performed by 8 subjects in a controlled environment with accelerometer sensors, motion capture and indoor location sensor. In this paper, we summarize the results of the competition.

    DOI: 10.1145/3341162.3345577

    Scopus

    Other Link: https://www.scopus.com/inward/record.uri?partnerID=HzOxMe3b&scp=85072882944&origin=inward

  • Reduction of marker-body matching work in activity recognition using motion capture Reviewed

    Takeda S., Lago P., Okita T., Inoue S.

    UbiComp/ISWC 2019- - Adjunct Proceedings of the 2019 ACM International Joint Conference on Pervasive and Ubiquitous Computing and Proceedings of the 2019 ACM International Symposium on Wearable Computers   835 - 842   2019.09

     More details

    Language:English   Publishing type:Research paper (international conference proceedings)

    In this paper, activity recognition is performed using an optical motion capture system that can measure three-dimensional position information of reflective markers attached to the body. The individual markers detected by motion capture are automatically associated with which part of the body they are attached to. However, due to the overlapping of obstacles and other body parts and misplacement of the markers, these may be hidden from the camera and enter a blind spot, which may frequently cause a marker to be associated to different body parts erroneously. Usually, these errors need to be corrected manually after measurement, but this work is very time consuming, cumbersome and requires some skill. In this research, it is thought that there is no problem in recognizing the activity even if the process of spending the effort of correcting the correspondence between the marker after measurement and the body is omitted in the activity recognition using the motion capture. Because feature quantities are extracted from activity data when performing action recognition, even if an error occurs in part of the marker data, the effect is small because the correct feature quantities are selected and other marker data can compensate for an error. In addition, in this paper, we proposed a method to recognize the activity using the data when the human body template preparation required before Mocap data measurement is omitted, which is one of marker body matching work. The verification showed that even if the marker body matching operation was omitted, it was possible to recognize the action with high accuracy.

    DOI: 10.1145/3341162.3345591

    Scopus

    Other Link: https://www.scopus.com/inward/record.uri?partnerID=HzOxMe3b&scp=85072894653&origin=inward

  • MEASURed: Evaluating Sensor-Based Activity Recognition Scenarios Reviewed International journal

    Paula Lago, Shingo Takeda, Tsuyoshi Okita, Sozo Inoue

    Human Activity Sensing: Corpus and Applications ( Springer Nature )   2019.09

     More details

    Language:English   Publishing type:Research paper (scientific journal)

    Human Activity Recognition from accelerometer sensors is key to enable applications such as fitness tracking or health status monitoring at home. However, evaluating the performance of activity recognition systems in real-life deployments is challenging to the multiple differences in sensor number, placement and orientation that may arise in real settings. Considering such differences requires a large amount of labeled data. To overcome the challenges and costs associated to the collection of a wide range of heterogeneous data, we propose a simulator, called MEASURed, which uses motion capture to simulate accelerometer data on different settings. Then, using the simulated data to estimate the performance of activity recognition models under different scenarios. In this chapter, we describe MEASURed and evaluate its performance in estimating the accuracy of activity recognition models. Our results show that MEASURed can estimate the average accuracy of an activity recognition model using real accelerometer magnitude data. By using motion capture to simulate accelerometer data, the sensor research community can profit from visual datasets that have been collected by other communities to evaluate performance of activity recognition in a wide range of activities. MEASURed can be used to evaluate activity recognition classifiers in settings with different number, placement, and sampling rate of accelerometer sensors. The evaluation on a broad spectrum of scenarios gives a more general view of models and their limitations.

  • Evaluation of transfer learning for human activity recognition among different datasets Reviewed

    Islam M.S., Okita T., Inoue S.

    Proceedings - IEEE 17th International Conference on Dependable, Autonomic and Secure Computing, IEEE 17th International Conference on Pervasive Intelligence and Computing, IEEE 5th International Conference on Cloud and Big Data Computing, 4th Cyber Science and Technology Congress, DASC-PiCom-CBDCom-CyberSciTech 2019   854 - 859   2019.08

     More details

    Language:English   Publishing type:Research paper (international conference proceedings)

    Human activity recognition is a potential area of research. For better performance, it requires significant amount of labelled data. Collecting labeled activity data is expensive and time-consuming. To solve this problem, transfer learning has been demonstrated very effective as it gathers knowledge from labeled train data of source domain and transfers that knowledge to target domain, which has little or no labeled data. In this paper, we propose unsupervised transfer learning from source dataset to target dataset, which are completely different in terms of number of users and samples. We have used Maximum Mean Discrepancy (MMD) based transfer learning model and compared with base Convolutional Neural Network (CNN) model. We have used 4 datasets for experiment. We have trained the model on a source dataset and then transferred the model to a target dataset, which has no labels to classify activities. We have found that transfer learning model has achieved better performance compared to the base model.

    DOI: 10.1109/DASC/PiCom/CBDCom/CyberSciTech.2019.00155

    Scopus

    Other Link: https://www.scopus.com/inward/record.uri?partnerID=HzOxMe3b&scp=85075116858&origin=inward

  • モーションキャプチャを用いた行動認識におけるマーカー身体対応付け作業の削減

    Shingo Takeda, Paula Lago, Tsuyoshi Okita, Sozo Inoue, Yoshinori Ideno

    DICOMOシンポジウムプロシーディングス   2019.07

     More details

    Language:Japanese   Publishing type:Research paper (conference, symposium, etc.)

  • センサ行動認識における機械学習のための実験室高精度データと現場長時間データの比較

    Hiroki Goto, Shingo Takeda, Paula Lago, Tsuyoshi Okita, Sozo Inoue

    DICOMOシンポジウムプロシーディングス   2019.07

     More details

    Language:Japanese   Publishing type:Research paper (conference, symposium, etc.)

  • 医療相談のための対話システムにおける要求分析

    二宮仁志, Tittaya Mairittha, 大北剛, 井上創造

    ソフト九州プロシーディングス   2018.12

     More details

    Language:Japanese   Publishing type:Research paper (conference, symposium, etc.)

  • センサ行動認識のための機械学習を用いた加速度データシミュレーション

    武田 紳吾, Paula Lago, 大北剛, 井上創造

    ソフト九州プロシーディングス   2018.12

     More details

    Language:Japanese   Publishing type:Research paper (conference, symposium, etc.)

  • 種々の敵対的生成ネットワークに対するGeometry Scoreによる評価

    福島康太, 大北剛, 井上創造

    ソフト九州プロシーディングス   2018.12

     More details

    Language:Japanese   Publishing type:Research paper (conference, symposium, etc.)

  • Summary of the Sussex-Huawei locomotion-transportation recognition challenge Reviewed

    Wang L., Murao K., Gjoreski H., Okita T., Roggen D.

    UbiComp/ISWC 2018 - Adjunct Proceedings of the 2018 ACM International Joint Conference on Pervasive and Ubiquitous Computing and Proceedings of the 2018 ACM International Symposium on Wearable Computers   1521 - 1530   2018.10

     More details

    Language:English   Publishing type:Research paper (international conference proceedings)

    In this paper we summarize the contributions of participants to the Sussex-Huawei Transportation-Locomotion (SHL) Recognition Challenge organized at the HASCA Workshop of UbiComp 2018. The SHL challenge is a machine learning and data science competition, which aims to recognize eight transportation activities (Still, Walk, Run, Bike, Bus, Car, Train, Subway) from the inertial and pressure sensor data of a smartphone. We introduce the dataset used in the challenge and the protocol for the competition. We present a meta-analysis of the contributions from 19 submissions, their approaches, the software tools used, computational cost and the achieved results. Overall, two entries achieved F1 scores above 90%, eight with F1 scores between 80% and 90%, and nine between 50% and 80%.

    DOI: 10.1145/3267305.3267519

    Scopus

    Other Link: https://www.scopus.com/inward/record.uri?partnerID=HzOxMe3b&scp=85058337765&origin=inward

  • 6th international workshop on human Activity Sensing Corpus and Applications (HASCA) Reviewed

    Murao K., Enokibori Y., Gjoreski H., Lago P., Okita T., Siirtola P., Hiroi K., Scholl P.M.

    UbiComp/ISWC 2018 - Adjunct Proceedings of the 2018 ACM International Joint Conference on Pervasive and Ubiquitous Computing and Proceedings of the 2018 ACM International Symposium on Wearable Computers   1392 - 1395   2018.10

     More details

    Language:English   Publishing type:Research paper (international conference proceedings)

    The recognition of complex and subtle human behaviors from wearable sensors will enable next-generation human-oriented computing in scenarios of high societal value (e.g., dementia care). This will require large-scale human activity corpuses and much improved methods to recognize activities and the context in which they occur. This workshop deals with the challenges of designing reproducible experimental setups, running large-scale dataset collection campaigns, designing activity and context recognition methods that are robust and adaptive, and evaluating systems in the real world. We wish to reflect on future methods, such as lifelong learning approaches that allow open-ended activity recognition. Unique this year, HASCA will welcome papers from participants to the Sussex-Huawei Locomotion and Transportation Recognition Competition in a special session.

    DOI: 10.1145/3267305.3274145

    Scopus

    Other Link: https://www.scopus.com/inward/record.uri?partnerID=HzOxMe3b&scp=85058343540&origin=inward

  • A multi-sensor setting activity recognition simulation tool Reviewed

    Takeda S., Okita T., Lago P., Inoue S.

    UbiComp/ISWC 2018 - Adjunct Proceedings of the 2018 ACM International Joint Conference on Pervasive and Ubiquitous Computing and Proceedings of the 2018 ACM International Symposium on Wearable Computers   1444 - 1448   2018.10

     More details

    Language:English   Publishing type:Research paper (international conference proceedings)

    Motion capture generates data which are often more accurate than those captured by multiple of accelerometer sensors by their physical specification. Based on the observation that accelerometer data can be obtained by the second derivation of position data from motion capture, we propose a simulator, called MEASURed, for activity recognition classifiers. MEASURed can accommodate any number of virtual accelerometer sensors on the body based on some given motion capture data. Therefore, MEASURed can evaluate activity recognition classifiers in settings with different number, placement, and sampling rate of accelerometer sensors. Our results show that the F1-Score estimated by MEASURed is close to that obtained with the real accelerometer data.

    DOI: 10.1145/3267305.3267509

    Scopus

    Other Link: https://www.scopus.com/inward/record.uri?partnerID=HzOxMe3b&scp=85058289678&origin=inward

  • Activity recognition: Translation across sensor modalities using deep learning Reviewed

    Okita T., Inoue S.

    UbiComp/ISWC 2018 - Adjunct Proceedings of the 2018 ACM International Joint Conference on Pervasive and Ubiquitous Computing and Proceedings of the 2018 ACM International Symposium on Wearable Computers   1462 - 1471   2018.10

     More details

    Language:English   Publishing type:Research paper (international conference proceedings)

    We propose a method to translate between multi-modalities using an RNN encoder-decoder model. Based on such a model allowing to translate between modalities, we built an activity recognition system. The idea of equivalence of modality was investigated by Banos et al. This paper replaces this with deep learning. We compare the performance of translation with/without clustering and sliding window. We show the preliminary performance of activity recognition attained the F1 score of 0.78.

    DOI: 10.1145/3267305.3267512

    Scopus

    Other Link: https://www.scopus.com/inward/record.uri?partnerID=HzOxMe3b&scp=85058312644&origin=inward

  • Poster: Improving sensor-based activity recognition using motion capture as additional information Reviewed

    Lago P., Okita T., Takeda S., Inoue S.

    UbiComp/ISWC 2018 - Adjunct Proceedings of the 2018 ACM International Joint Conference on Pervasive and Ubiquitous Computing and Proceedings of the 2018 ACM International Symposium on Wearable Computers   118 - 121   2018.10

     More details

    Language:English   Publishing type:Research paper (international conference proceedings)

    We propose a new method for human activity recognition using a single accelerometer sensor and additional sensors for training. The performance of inertial sensors for complex activities drops considerably compared with simple activities due to inter-class similarities. In such cases deploying more sensors may improve the performance. But such strategy is often not feasible in reality due to costs or privacy concerns among others. In this context, we propose a new method to use additional sensors only in training phase. We introduce the idea of mapping the test data to a codebook created from the additional sensor information. Using the Berkeley MHAD dataset our preliminary results show this worked positively; improving in 10.0% both the average F1-score and the average accuracy. Notably, the improvement for the stand, sit and sit to stand activities was higher, typical activities for which the inertial sensor is less informative when using the wrist-worn accelerometer.

    DOI: 10.1145/3267305.3267596

    Scopus

    Other Link: https://www.scopus.com/inward/record.uri?partnerID=HzOxMe3b&scp=85058340955&origin=inward

  • Pre-consulting dialogue systems for telemedicine: Yes/no intent classification Reviewed

    Mairittha T., Okita T., Inoue S.

    UbiComp/ISWC 2018 - Adjunct Proceedings of the 2018 ACM International Joint Conference on Pervasive and Ubiquitous Computing and Proceedings of the 2018 ACM International Symposium on Wearable Computers   742 - 745   2018.10

     More details

    Language:English   Publishing type:Research paper (international conference proceedings)

    Telemedicine is an emerging challenge for the shortage of qualified professionals, particularly in under-resourced regions. Physical assessment by a non-medical doctor is a practice in telemedicine which discovers essential symptom of a patient who needs to consult a doctor. We aim at facilitating this stage with a conversational chatbot which identifies the patient by conversation. Adopting the procedures of physical assessment one critical types of conversation involves in the self-diagnosis. Further, it turned out that useful kinds of questions in chatbot at this stage are related to Yes/No questions. We discovered that particular difficulties lie in the ambiguous replies by the patients: a patient modifies a question which makes them answer yes or no, a response does not the corresponding reply to the question, a reply involves some part yes and some part no, and so on. Focusing on this particular type of question we introduce a text classifier using Long Short-Term Memory (LSTM) and build a corpus using Twitter.

    DOI: 10.1145/3267305.3267704

    Scopus

    Other Link: https://www.scopus.com/inward/record.uri?partnerID=HzOxMe3b&scp=85058328754&origin=inward

  • Dialogue breakdown detection with long short term memory Reviewed

    Mairittha T., Okita T., Inoue S.

    Lecture Notes of the Institute for Computer Sciences, Social-Informatics and Telecommunications Engineering, LNICST   240   245 - 250   2018.01

     More details

    Language:English   Publishing type:Research paper (international conference proceedings)

    This paper aims to detect the utterance which can be categorized as the breakdown of the dialogue flow. We propose a logistic regression-based and a Long Short-Term Memory (LSTM)-based methods. Using the input with utterance-response pairs, the performance of the LSTM-based method is superior to that of the logistic regression-based method in 36% measured with F-measure. We also measured the performance using the performance with utterance-response pairs: the performance with the input only with responses is unexpectedly inferior to those with responses in 6% to 23% measured with F-measure.

    DOI: 10.1007/978-3-319-90740-6_18

    Scopus

    Other Link: https://www.scopus.com/inward/record.uri?partnerID=HzOxMe3b&scp=85046867198&origin=inward

  • ダンスの上手い人のマイニング的な分析

    OKITA Tsuyoshi, INOUE Sozo

    JSAI Technical Report, Type 2 SIG ( The Japanese Society for Artificial Intelligence )   2018 ( 18 )   03   2018.01

     More details

    Authorship:Lead author   Language:English   Publishing type:Research paper (conference, symposium, etc.)

    DOI: 10.11517/jsaisigtwo.2018.AM-18_03

    CiNii Article

    CiNii Research

    Other Link: https://ci.nii.ac.jp/naid/130008080212

  • Parallelization of Neural Network Training for NLP with Hogwild! Reviewed International journal

    Valentin Deyringer and Alexander Fraser and Helmut Schmid and Tsuyoshi Okita

    The Prague Bulletin of Mathematical Linguistics ( Charles University, Czech Republic )   109   29 - 38   2017.10

     More details

    Language:English   Publishing type:Research paper (scientific journal)

    DOI: 10.1515/pralin-2017-0036

  • Recognition of multiple overlapping activities using compositional cnn-lstm model Reviewed

    Okita T., Inoue S.

    UbiComp/ISWC 2017 - Adjunct Proceedings of the 2017 ACM International Joint Conference on Pervasive and Ubiquitous Computing and Proceedings of the 2017 ACM International Symposium on Wearable Computers   165 - 168   2017.09

     More details

    Language:English   Publishing type:Research paper (international conference proceedings)

    This paper introduces a new task, a recognition of multiple overlapping activities in the context of activity recognition. We propose a compositional CNN+LSTM algorithm. The experimental results show on the artificial dataset that it improved the accuracy from 27% to 43%.

    DOI: 10.1145/3123024.3123095

    Scopus

    Other Link: https://www.scopus.com/inward/record.uri?partnerID=HzOxMe3b&scp=85030836480&origin=inward

  • The DCU discourse parser for connective, argument identification and explicit sense classification Reviewed International journal

    Wang L., Hokamp C., Okita T., Zhang X., Liu Q.

    CoNLL 2015 - 19th Conference on Computational Natural Language Learning, Proceedings of the Shared Task   89 - 94   2015.01

     More details

    Language:English   Publishing type:Research paper (international conference proceedings)

    This paper describes our submission to the CoNLL-2015 shared task on discourse parsing. We factor the pipeline into sub-components which are then used to form the final sequential architecture. Focusing on achieving good performance when inferring explicit discourse relations, we apply maximum entropy and recurrent neural networks to different sub-tasks such as connective identification, argument extraction, and sense classification. The our final system achieves 16.51%, 12.73% and 11.15% overall F1 scores on the dev, WSJ and blind test sets, respectively.

    Scopus

    Other Link: https://www.scopus.com/inward/record.uri?partnerID=HzOxMe3b&scp=84973379014&origin=inward

  • The DCU discourse parser: A sense classification task Reviewed International journal

    Okita T., Wang L., Liu Q.

    CoNLL 2015 - 19th Conference on Computational Natural Language Learning, Proceedings of the Shared Task   71 - 77   2015.01

     More details

    Language:English   Publishing type:Research paper (international conference proceedings)

    This paper describes the discourse parsing system developed at Dublin City University for participation in the CoNLL 2015 shared task. We participated in two tasks: a connective and argument identification task and a sense classification task. This paper focuses on the latter task and especially the sense classification for implicit connectives.

    Scopus

    Other Link: https://www.scopus.com/inward/record.uri?partnerID=HzOxMe3b&scp=85072752049&origin=inward

  • The DCU Terminology Translation System for the Medical Query Subtask at WMT14 Reviewed International journal

    Xiaofeng Wu, Rejwanul Haque, Tsuyoshi Okita, Piyush Arora, Andy Way, Qun Liu

    Proceedings of the 9th Workshop on Statistical Machine Translation   2014.06

     More details

    Language:English   Publishing type:Research paper (international conference proceedings)

  • Active Learning-based Local Graph Matching for Textual Entailment Reviewed International journal

    Tsuyoshi Okita

    Proceedings of the 10th International Symposium on Natural Language Processing   2013.10

     More details

    Authorship:Lead author   Language:English   Publishing type:Research paper (international conference proceedings)

  • Shallow Semantically-Informed PBSMT and HPBSMT Reviewed International journal

    Tsuyoshi Okita, Qun Liu, Josef van Genabith

    Proceedings of the 8th Workshop on Statistical Machine Translation   2013.08

     More details

    Authorship:Lead author   Language:English   Publishing type:Research paper (international conference proceedings)

  • Joint Space Neural Probabilistic Language Model for Statistical Machine Translation International journal

    Tsuyoshi Okita

    arxiv   2013.01

     More details

    Authorship:Lead author   Language:English   Publishing type:Research paper (bulletin of university, research institution)

    arXiv

  • Results from the ML4HMT-12 Shared Task on Applying Machine Learning Techniques to Optimise the Division of Labour in Hybrid Machine Translation Reviewed International journal

    Christian Federmann, Tsuyoshi Okita, Maite Melero, Marta R. Costa-Jussa, Toni Badia and Josef van Genabith

    Proceedings of ML4HMT Workshop   2012.12

     More details

    Language:English   Publishing type:Research paper (international conference proceedings)

  • Neural Probabilistic Language Model for System Combination Reviewed International journal

    Tsuyoshi Okita

    Proceedings of ML4HMT Workshop   2012.12

     More details

    Authorship:Lead author   Language:English   Publishing type:Research paper (international conference proceedings)

  • Topic Modeling-based Domain Adaptation for System Combination Reviewed International journal

    Tsuyoshi Okita and Antonio Toral and Josef van Genabith

    Proceedings of ML4HMT Workshop   2012.12

     More details

    Authorship:Lead author   Language:English   Publishing type:Research paper (international conference proceedings)

  • Sentence-level Quality Estimation for MT System Combination Reviewed International journal

    Tsuyoshi Okita and Raphael Rubino and Josef van Genabith

    Proceedings of ML4HMT Workshop   2012.12

     More details

    Authorship:Lead author   Language:English   Publishing type:Research paper (international conference proceedings)

  • System Combination with Extra Alignment Information Reviewed International journal

    Xiaofeng Wu, Tsuyoshi Okita, Josef van Genabith, and Qun Liu

    Proceedings of ML4HMT Workshop   2012.12

     More details

    Language:English   Publishing type:Research paper (international conference proceedings)

  • Minimum Bayes risk decoding with enlarged hypothesis space in system combination Reviewed

    Okita T., Van Genabith J.

    Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)   7182 LNCS ( PART 2 )   40 - 51   2012.03

     More details

    Language:English   Publishing type:Research paper (international conference proceedings)

    This paper describes a new system combination strategy in Statistical Machine Translation. Tromble et al. (2008) introduced the evidence space into Minimum Bayes Risk decoding in order to quantify the relative performance within lattice or n-best output with regard to the 1-best output. In contrast, our approach is to enlarge the hypothesis space in order to incorporate the combinatorial nature of MBR decoding. In this setting, we perform experiments on three language pairs ES-EN, FR-EN and JP-EN. For ES-EN JRC-Acquis our approach shows 0.50 BLEU points absolute and 1.9% relative improvement obver the standard confusion network-based system combination without hypothesis expansion, and 2.16 BLEU points absolute and 9.2% relative improvement compared to the single best system. For JP-EN NTCIR-8 the improvement is 0.94 points absolute and 3.4% relative, and for FR-EN WMT09 0.30 points absolute and 1.3% relative compared to the single best system, respectively. © 2012 Springer-Verlag.

    DOI: 10.1007/978-3-642-28601-8_4

    Scopus

    Other Link: https://www.scopus.com/inward/record.uri?partnerID=HzOxMe3b&scp=84858307159&origin=inward

  • Annotated corpora for word alignment between Japanese and english and its evaluation with MAP-based word aligner Reviewed

    Okita T.

    Proceedings of the 8th International Conference on Language Resources and Evaluation, LREC 2012   3241 - 3248   2012.01

     More details

    Language:English   Publishing type:Research paper (international conference proceedings)

    This paper presents two annotated corpora for word alignment between Japanese and English. We annotated on top of the IWSLT-2006 and the NTCIR-8 corpora. The IWSLT-2006 corpus is in the domain of travel conversation while the NTCIR-8 corpus is in the domain of patent. We annotated the first 500 sentence pairs from the IWSLT-2006 corpus and the first 100 sentence pairs from the NTCIR-8 corpus. After mentioned the annotation guideline, we present two evaluation algorithms how to use such hand-annotated corpora: although one is a well-known algorithm for word alignment researchers, one is novel which intends to evaluate a MAP-based word aligner of Okita et al. (2010b).

    Scopus

    Other Link: https://www.scopus.com/inward/record.uri?partnerID=HzOxMe3b&scp=85037351924&origin=inward

  • Statistical machine translation with factored translation model: MWEs, separation of affixes, and others Reviewed

    Okita T., Ceausu A., Way A.

    Proceedings of the 24th International Florida Artificial Intelligence Research Society, FLAIRS - 24   353 - 354   2011.09

     More details

    Language:English   Publishing type:Research paper (international conference proceedings)

    This paper discusses Statistical Machine Translation when the target side is morphologically richer language. This paper intends to discuss the issues which are not covered by a factored translation model of Moses especially targetting EN-JP translation: the effect of Multi-Word Expressions, the separation of affixes, and other monolingual morphological issues. We intend to discuss these over a factored translation model. Copyright © 2011, Association for the Advancement of Artificial Intelligence. All rights reserved.

    Scopus

    Other Link: https://www.scopus.com/inward/record.uri?partnerID=HzOxMe3b&scp=80052409677&origin=inward

  • Given bilingual terminology in statistical machine translation: MWE-sensitive word alignment and hierarchical Pitman-Yor process-based translation model smoothing Reviewed

    Okita T., Way A.

    Proceedings of the 24th International Florida Artificial Intelligence Research Society, FLAIRS - 24   269 - 274   2011.09

     More details

    Language:English   Publishing type:Research paper (international conference proceedings)

    This paper considers a scenario when we are given almost perfect knowledge about bilingual terminology in terms of a test corpus in Statistical Machine Translation (SMT). When the given terminology is part of a training corpus, one natural strategy in SMT is to use the trained translation model ignoring the given terminology. Then, two questions arises here. 1) Can a word aligner capture the given terminology? This is since even if the terminology is in a training corpus, it is often the case that a resulted translation model may not include these terminology. 2) Are probabilities in a translation model correctly calculated? In order to answer these questions, we did experiment introducing a Multi-Word Expression-sensitive (MWE-sensitive) word aligner and a hierarchical Pitman-Yor process-based translation model smoothing. Using 200k JP-EN NTCIR corpus, our experimental results show that if we introduce an MWE-sensitive word aligner and a new translation model smoothing, the overall improvement was 1.35 BLEU point absolute and 6.0% relative compared to the case we do not introduce these two. Copyright © 2011, Association for the Advancement of Artificial Intelligence. All rights reserved.

    Scopus

    Other Link: https://www.scopus.com/inward/record.uri?partnerID=HzOxMe3b&scp=80052406399&origin=inward

  • Pitman-Yor Process-based Language Mode Reviewed International journal

    Tsuyoshi Okita and Andy Way

    International Journal of Asian Language Processing   21 ( 2 )   55 - 70   2011.04

     More details

    Authorship:Lead author   Language:English   Publishing type:Research paper (scientific journal)

  • Pitman-Yor Process-based Language Model Reviewed International journal

    Tsuyoshi Okita, Andy Way

    International Journal of Asian Language Processing   21 ( 2 )   57 - 70   2011.03

     More details

    Authorship:Lead author   Language:English   Publishing type:Research paper (scientific journal)

  • MWE-sensitive Word Aligner in Factored Translation Model International journal

    Tsuyoshi Okita, Andy Way

    Proceedings of Machine Translation and Morphologically-rich Languages   2011.01

     More details

    Authorship:Lead author   Language:English   Publishing type:Research paper (conference, symposium, etc.)

    Israel   Haifa   2011.01.23  -  2011.01.27

  • Hierarchical Pitman-Yor language model for machine translation Reviewed

    Okita T., Way A.

    Proceedings - 2010 International Conference on Asian Language Processing, IALP 2010   245 - 248   2010.12

     More details

    Language:English   Publishing type:Research paper (international conference proceedings)

    The hierarchical Pitman-Yor process-based smoothing method applied to language model was proposed by Goldwater and by Teh; the performance of this smoothing method is shown comparable with the modified Kneser-Ney method in terms of perplexity. Although this method was presented four years ago, there has been no paper which reports that this language model indeed improves translation quality in the context of Machine Translation (MT). This is important for the MT community since an improvement in perplexity does not always lead to an improvement in BLEU score; for example, the success of word alignment measured by Alignment Error Rate (AER) does not often lead to an improvement in BLEU. This paper reports in the context of MT that an improvement in perplexity really leads to an improvement in BLEU score. It turned out that an application of the Hierarchical Pitman-Yor Language Model (HPYLM) requires a minor change in the conventional decoding process. Additionally to this, we propose a new Pitman-Yor process-based statistical smoothing method similar to the Good-Turing method although the performance of this is inferior to HPYLM. We conducted experiments; HPYLM improved by 1.03 BLEU points absolute and 6% relative for 50k EN-JP, which was statistically significant. © 2010 IEEE.

    DOI: 10.1109/IALP.2010.34

    Scopus

    Other Link: https://www.scopus.com/inward/record.uri?partnerID=HzOxMe3b&scp=79551520331&origin=inward

  • Gap Between Theory and Practice: Noise Sensitive Word Alignment in Machine Translation Reviewed International journal

    Tsuyoshi Okita, Yvette Graham, Andy Way

    Journal of Machine Learning Research Workshop and Conference Proceedings   11   119 - 126   2010.09

     More details

    Authorship:Lead author   Language:English   Publishing type:Research paper (scientific journal)

  • Multi-Word Expression-Sensitive Word Alignment Reviewed International journal

    Tsuyoshi Okita, Alfredo Maldonado Guerra, Yvette Graham, Andy Way

    In Proceedings of the Fourth International Workshop On Cross Lingual Information Access ( Coling 2010 Organizing Committee )   26 - 33   2010.08

     More details

    Authorship:Lead author   Language:English   Publishing type:Research paper (scientific journal)

    China   Beijing   2010.08.28  -  2010.08.28

    This paper presents a new word align-
    ment method which incorporates knowl-
    edge about Bilingual Multi-Word Expres-
    sions (BMWEs). Our method of word
    alignment first extracts such BMWEs in
    a bidirectional way for a given corpus and
    then starts conventional word alignment,
    considering the properties of BMWEs in
    their grouping as well as their alignment
    links. We give partial annotation of align-
    ment links as prior knowledge to the word
    alignment process; by replacing the max-
    imum likelihood estimate in the M-step
    of the IBM Models with the Maximum A
    Posteriori (MAP) estimate, prior knowl-
    edge about BMWEs is embedded in the
    prior in this MAP estimate. In our exper-
    iments, we saw an improvement of 0.77
    Bleu points absolute in JP–EN. Except
    for one case, our method gave better re-
    sults than the method using only BMWEs
    grouping. Even though this paper does
    not directly address the issues in Cross-
    Lingual Information Retrieval (CLIR), it
    discusses an approach of direct relevance
    to the field. This approach could be
    viewed as the opposite of current trends
    in CLIR on semantic space that incorpo-
    rate a notion of order in the bag-of-words
    model (e.g. co-occurences).

    Kyutacar

  • Data cleaning for word alignment Reviewed

    Okita T.

    ACL-IJCNLP 2009 - Joint Conf. of the 47th Annual Meeting of the Association for Computational Linguistics and 4th Int. Joint Conf. on Natural Language Processing of the AFNLP, Proceedings of the Conf.   72 - 80   2009.01

     More details

    Language:English   Publishing type:Research paper (international conference proceedings)

    Parallel corpora are made by human beings. However, as an MT system is an aggregation of state-of-the-art NLP technologies without any intervention of human beings, it is unavoidable that quite a few sentence pairs are beyond its analysis and that will therefore not contribute to the system. Furthermore, they in turn may act against our objectives to make the overall performance worse. Possible unfavorable items are n : m mapping objects, such as paraphrases, non-literal translations, and multiword expressions. This paper presents a pre-processing method which detects such unfavorable items before supplying them to the word aligner under the assumption that their frequency is low, such as below 5 percent. We show an improvement of Bleu score from 28.0 to 31.4 in English-Spanish and from 16.9 to 22.1 in German-English. © 2009 ACL and AFNLP.

    DOI: 10.3115/1667884.1667895

    Scopus

    Other Link: https://www.scopus.com/inward/record.uri?partnerID=HzOxMe3b&scp=80052416271&origin=inward

  • Low-Resource Machine Translation Using MATREX: The DCU Machine Translation System for IWSLT 2009 Reviewed

    Ma Y., Okita T., Çetinoǧlu Ö., Du J., Way A.

    2009 International Workshop on Spoken Language Translation, IWSLT 2009   29 - 36   2009.01

     More details

    Language:English   Publishing type:Research paper (international conference proceedings)

    Scopus

    Other Link: https://www.scopus.com/inward/record.uri?partnerID=HzOxMe3b&scp=85133165029&origin=inward

  • Load Balance Protocol of Cluster on Grid: Pervasive Maximum Algorithmic Parallelism Reviewed International journal

    Tsuyoshi Okita

    Procedings of the 6th International Conference on Principles of Distributed Systems (OPODIS 2002) ( Suger, Saint-Denis, rue Catulienne, France )   3   203 - 210   2002.12

     More details

    Authorship:Lead author   Language:English   Publishing type:Research paper (international conference proceedings)

    France   Reims   2002.12.11  -  2002.12.13

    Pervasive computing software help to manage information and reduce the complexity of available computing resources in a timely manner. On the other hand, a grid is a resource whose information is changing anytime and anywhere, where the availability of CPUs is only informed in run time when a cluster asks to the grid. On the other hand, a cluster is often implemented in a static way, which assumes some particular parallel architecture and the number of CPUs. Even when a cluster can consume maximum available CPU resources, if a cluster i implemented assuming the number of CPUs, a cluster could not run using more than this number of CPUs. Our mechanim of load balance protocol of cluster provides a dynamic way of implementing a cluster that can consume maximum available CPU resources on run time. In order to do so, we propoe two mechanisms: to provide yet another (graphical) parallel language to describe clusters and to provide the protocols between a cluster on a grid. While many paralell languages resolve parallel architecture dependencies in compile time, our parallel language resolve parallel architecture dependencies in run time. Our load balance protocol bases on this (graphical) parallel language and it provides implementation of them.

    Kyutacar

  • PRTccp: Priority-driven Real-Time Concurrent Constraint Programming Reviewed International journal

    Tsuyoshi Okita

    Proceedings of the 14th Nordic Workshop on Programming Theory   2002.11

     More details

    Authorship:Lead author   Language:English   Publishing type:Research paper (international conference proceedings)

    Estonia   Tallinn   2002.11.20  -  2002.11.22

    This paper presents real-time formal language PRTccp, which complements priority-driven concerns in Tccp. First, we showed the necessity of formal language for priority-driven system compared to reactive real-time system. Secondly, we showed the grammar of PRTccp. Thirdly, we showed a small example of scheduler.

    Kyutacar

▼display all

Conference Prsentations (Oral, Poster)

  • JPEGの画像表現を用いた画像生成の高速化

    大北剛, 管谷克彦, 坂本比呂志

    IBIS 2021 

     More details

    Event date: 2021.11.10 - 2021.11.12   Language:English  

  • グローバルな情報を加味するセマンティックセグメンテーションとラベルの重複を許す分類のジョイント学習

    平野北斗, 竹本和広, 大北剛

    第23回情報理論的学習理論ワークショップ(IBIS2020) 

     More details

    Event date: 2020.11.23 - 2020.11.26   Language:English  

  • Spatio-temporal Model for Intracerebral Hemorrhage: Embedding Methods Solving Violating Assumptions on ML

    Tsuyoshi Okita

    LRML Learning Meaningful Representations of Life Workshop 

     More details

    Event date: 2019.12.13   Language:English  

  • 欠損値問題や不均衡データを埋め込みとして用いた時空間機械学習モデル

    大北剛

    欠損値問題や不均衡データを埋め込みとして用いた時空間機械学習モデル  IBIS2019

     More details

    Event date: 2019.11.20 - 2019.11.23   Language:Japanese  

  • Translation of Signals: Wave2wave

    Tsuyoshi Okita, Hirotaka Hachiya, Sozo Inoue, Naonori Ueda

    Discovery Science 

     More details

    Event date: 2019.10.28 - 2019.10.30   Language:English  

  • メビウス型包除積分ニューラルネットワークによるデータ解析

    本田あおい, 大北剛

    実解析学シンポジウム 

     More details

    Event date: 2019.10.25 - 2019.10.27   Language:Japanese  

  • Cross Modal Translation of Signals and Its Application to Activity Recognition Invited

    Tsuyoshi Okita, Hirotaka Hachiya, Sozo Inoue, Naonori Ueda

    The First Japan-Israel Machine Learning Workshop 

     More details

    Event date: 2018.11.19 - 2018.11.20   Language:English  

  • Language Grounded Planning for Dialogue Systems

    Tsuyoshi Okita and Sozo Inoue

    NVIDIA's GPU Technology Conference (GTC)  

     More details

    Event date: 2018.09.13 - 2018.09.14   Language:English  

  • Language Grounded Activity Recognition and Planning

    Tsuyoshi Okita and Sozo Inoue

    Second International Workshop on Symbolic-Neural Learning (SNL-2018) 

     More details

    Event date: 2018.07.05 - 2018.07.06   Language:English  

  • Brazilator: Machine Translation and Sentiment Analysis for World Cup 2014

    Santiago Cortes, Piyush Arora, Chris Hokamp, Federico Fancellu, Alex Killen, Lamia Tounsi, Antonio Toral, Ankit Srivastava, Maria Alecu, Iacer Calixto, Sheila Castilho, Keith Curtis, Federico Gaspari, Akira Hayakawa, Teresa Lynn, Peyman Passban, Eziz Tursun, Ali Hosseinzadeh Vahid, Xiaofeng Wu, Xiaojun Zhang, Debasis Ganguly, Louise Irwin, Anna Kostekidou, Liangyou Li, Tsuyoshi Okita, Ximo Planells, David Racca, Joris Vreeke, Jian Zhang, Andy Way, Will Lewis, Declan Groves, Federico Garcea, and Chris Wendt

    Association for Machine Translation in the Americas  AMTA

     More details

    Event date: 2014.10.22 - 2014.10.26   Language:English  

    デモ発表

▼display all

Industrial Property

  • データ予測装置及びデータ予測方法

    大北 剛、井上 創造

     More details

    Application no:特開2020-61145(P2020-61145A)  Date applied:2019.10.04

    ディープラーニング(深層学習)を用いて異なるモダリティ間でのデータの変換を行うことにより、各種のモダリティを有効活用して、その活用範囲を広げることができるデータ予測装置及びデータ予測方法を提供する。

  • 計表処理装置および計表処理方法、並びに伝送媒体

    大北 剛

     More details

    Application no:特願平9−283310  Date applied:1997.10.16

    Announcement no:特開平11−120257  Date announced:1999.04.30

    【課題】計表の行側の項目および列側の項目を、階層との整合を図りながら作成することができるようにする。
    【解決手段】階層V−2を3個の項目に分割したとき、その上位の階層V−1は変更しないが、下位の階層V−3は、階層V−2に対応して3個の項目に分割する。行の項目と列の項目で規定される文字を入力する領域も、項目に対応して分割する。

Lectures

  • NeurIPS2020における深層学習の研究動向

    AIトレンドトップカンファレンス報告会  2021.03  人工知能学会

     More details

    Language:Japanese   Presentation type:Special lecture  

    NeurIPS2020における深層学習の研究動向をレポートした.

Honors and Awards

  • Outstanding Article Award --- 2021 Editor's Pick: Computer Science

    Frontiers   2022.01.28

    Lin Wang, Hristijan Gjoreski, Mathias Ciliberto, Paula Lago, Kazuya Murao, Tsuyoshi Okita, Daniel Roggen

     More details

    Country:Switzerland

    Dear Dr. Wang, Dr. Gjoreski, Dr. Ciliberto, Dr. Lago, Dr. Murao, Dr. Okita, and Dr. Roggen,

    It is with great pleasure that I inform you that your article "Three-Year Review of the 2018-2020 SHL Challenge on Transportation and Locomotion Mode Recognition From Mobile Sensors" was selected for our Outstanding Article award.

    The Chief Editors and the Editorial Office recognize the impact that your article is already having in the community as measured by early views and downloads. Congratulations!

    As such, we will highlight your article by featuring it in the dedicated collection "2021 Editor's Pick: Computer Science".

    Thanking you for your contributions to our journal and community, I wish you a fantastic 2022.

    Best regards,
    Rossana Isola
    Journal Manager

  • Best Paper Award

    Activity and Behavior Computing国際会議   2020.08.29

    Kohei Adachi, Paula Lago, Tsuyoshi Okita, Sozo Inoue

     More details

    Country:Japan

    Improvement of Human Action Recognition Using 3D Pose Estimation

Grants-in-Aid for Scientific Research

  • 計算モデルにガイドされた急成長を伴う時空間モデルの開発

    Grant number:20K12065  2020.04 - 2025.03   基盤研究(C)

Other External Funds

  • Precision medicineの確立に資する統合医療データベースの利活用に関する研究

    2019.04 - 2022.03

    厚生労働省 科学研究費補助金 政策科学総合研究事業  

Activities of Academic societies and Committees

  • DICOMOシンポジウム   プログラム委員  

    2023.02 - 2023.09

Social activity outside the university

  • データサイエンスプロ短期コース/深層学習特化型公開講座

    Role(s):Lecturer, Organizing member

    2023.03.14 - 2023.03.30

     More details

    Audience: Teachers, Researchesrs, General, Company

    Type:Other

  • データサイエンスプロ短期コース/深層学習特化型公開講座

    Role(s):Lecturer, Planner, Organizing member

    九州工業大学  データサイエンスプロ短期コース/深層学習特化型公開講座   2022.02.22 - 2022.03.31

     More details

    Audience: Researchesrs, General

    Type:Other

    全15コマの講義で, 参加者は33人. 165万円の収益を得た.

  • データサイエンスプロ短期コース/深層学習特化型公開講座

    Role(s):Lecturer, Organizing member

    九州工業大学  データサイエンスプロ短期コース/深層学習特化型公開講座  Virtual  2021.03.02 - 2021.03.30

     More details

    Audience: Researchesrs, General

    Type:Visiting lecture

    深層学習の講座全14コマ, 火木コースと土日コースの2回実施(累計28コマの講義)
    火木コース(3月2日,4日,9日,11日,16日,18日,23日,25日,30日)
    土日コース(3月6日,14日,21日,27日,28日)
    講師: 大北 剛, 井上 創造, 齊藤 剛史, 徳永 旭将, 竹本 和広

    18人が火曜木曜の平日コース, 5人が土曜日曜の週末コースであった.

    全体で23人X5万円=115万円を得た.

  • データサイエンスプロ短期コース/機械学習講座

    Role(s):Lecturer, Planner, Organizing member

    2019.04.01 - 2020.03.31

     More details

    Audience: Researchesrs, General

    Type:Visiting lecture

    機械学習講座をトヨタ自動車九州(宮田)にて行う. 3コマ全17回.